A Study Of Patients’ Willingness To Share Data On Personal Health Records

[Purpose] The personal health records（PHRs）are characterized by diversity,concentration and high quality due to the considerable amount of medical and routine data recorded.However,the data itself does not generate value.Its value lies in the analysis and utilization to help related investigation.Sharing is useful to data analysis and reuse,but the current situation of sharing is not optimistic.Therefore,the study on patients’ willingness to share data in PHRs is beneficial to understand their attitudes on sharing,which helps to provide reference for improving enthusiasm of sharing,improving utilization rate of PHRs and improving use value of health data.[Methods] The current situation and theories of relevant researches at home and abroad were sorted out through literature analysis,and the research hypothesis was put forward on this basis.Taking the patients of PHRs as the research object,the model of patients’ willingness to share was constructed from three aspects of perceived privacy risk,perceived benefit and altruism.Online questionnaire was sent to patient portal,health forum and patient chat group,and 400 valid sample data were collected.Confirmatory factor analysis（CFA）was used to verify reliability,convergent validity and discriminant validity of the questionnaire.It was reasonable to use descriptive statistical analysis to investigate sampling.Then the hypothesis model was verified,and the fitting index was calculated to verify the fitting degree of the model.AMOS was used to analyze the measurement model and structural model to verify the impact of variables in the basic hypothesis on the intention to share,as well as the antecedent impact of perceived benefits and perceived privacy risks.SPSS was used to perform hierarchical multivariate linear regression to verify the influence of control variables on shared intention.[Results] All numerical results of CFA reached the recommended values of previous scholars,indicating that the measurement questionnaire has good internal consistency,convergent validity and discriminant validity.In the verification of hypothesis model,the goodness of fit index reached the recommended value,indicating that the degree of fitting is acceptable.From results,it can be seen that patients’ willingness to share is influenced by perceived privacy risk,trust and altruism,among which altruism has the strongest influence.In the dimension of perceived privacy risk,privacy concern is the only independent variable that affects privacy risk,and it has a strong explanatory ability.Personal service and social support significantly affect perceived benefits.The influence of control variables on the shared intention is not significant.[Conclusions] This study makes an attempt to apply the Privacy Calculus Theory in the field of PHR.Altruism has the strongest effect on the willingness（p<0.001）,and one of the biggest concerns about privacy risks for patients is the disclosure of private information（p<0.001）,but more research is needed to confirm the possibility in health information field.In addition,personal service（p<0.001）and social support（p<0.001）significantly affect perceived benefits,indicating that patients accept and recognize the benefits of using PHR.It is suggested that we should pay attention to the protection of privacy information while encouraging patients to provide information.We should gradually establish norms of mutual trust and foster an atmosphere of social trust.PHR system in the future should make efforts from the directions of improving interface friendliness,strengthening doctor-patient communication,strengthening internal patient communication and other aspects which can promote personal service and social support,to encourage more patients to use and share PHR.[Innovations and Limitations] This study makes such innovations:（1）Privacy Calculus Theory is applied in the study of PHR and patients’ willingness to share data;（2）The antecedents of perceived benefits and perceived privacy risks are analyzed,and factors that encourage patients to disclose are considered.However,there are still following limitations:（1）There are ten hypotheses in this paper,four of which are untenable and the proportion is relatively high.Although explanations have been made for untenable hypothesis,the evidence is still insufficient;（2）The demographics of this survey sample are similar to the general population in some respects,but some differences may limit the prevalence of the survey results among individual health record users;（3）The discussion in this paper is limited to comparisons with previous studies that have raised the same or similar questions,which limits the ability to compare results with existing literature.